论文部分内容阅读
在临床前药物动力学研究中,通常无法获得每只动物个体的完整药物动力学观测数据。在某些情况下,每只动物个体只能提供一个观测数据。人们并不知道如何有效利用此类数据进行药物动力学参数的估算。本研究旨在比较一种新方法和传统方法在估算“单动物个体一单观测值”型数据的药物动力学参数时的优劣。本研究假设共有15只动物分别单次静脉注射相同剂量的药物,每只动物提供一个观测数据。共有5个观测时间点,每个观测时间点包括三个观测数据。数据仿真采用符合一级消除的一室模型。清除率(CL)和表观分布容积(V)的个体间变异均包含10%、30%和50%三个水平,统计模型选定为比例型残差模型,也包括三个水平:10%、30%和50%。本研究对比了药物动力学参数估算的两种方法(传统方法和有限重复抽样法),传统方法(M1)直接对原始数据(即“单动物个体一单观测值”型数据)进行参数估算,而有限重复抽样法(M2)则是按照观测时间点对原始数据进行排列组合,将其扩展为一套由含有完整药物动力学数据的虚拟动物个体组成的新的数据集,本研究中共243(C_3~1×C_3~1×C_3~1×C_3~1×C_3~1)只虚拟动物个体。本研究共重复了100次仿真,仿真与参数估算均采用NONMEM软件完成。结果显示,M2方法所估算的CL和v与其相应的仿真值更接近,但在不同ⅡV及残差的组合下稍有差异。总体而言,M2方法的优势随着残差的增大而减小,其也同样收到ⅡⅤ大小的影响,ⅡV增大时M2优势亦会下降。同M1方法类似,M2方法对参数的ⅡV也没有还原能力。与传统方法相比,有限重复抽样法在估算“单动物个体—单观测值”型数据药物动力学参数时可以提供更加可靠的结果。与个体间变异相比,估算结果主要收到残差大小的影响。
In preclinical pharmacokinetic studies, complete pharmacokinetic data are not usually available for each individual animal. In some cases, only one observation per animal is provided. People do not know how to effectively use such data for the estimation of pharmacokinetic parameters. The purpose of this study was to compare the advantages and disadvantages of a new method and a traditional method in estimating the pharmacokinetic parameters of a single observed single body of individuals. This study assumes that a total of 15 animals were given a single intravenous injection of the same dose of drug, each animal to provide an observation data. There are 5 observation time points, and each observation time point includes three observation data. The data simulations used a one-compartment model that matched the primary elimination. The inter-individual variability of the CL and the V contained 10%, 30% and 50% respectively. The statistical model was selected as the proportional residual model and included three levels: 10% , 30% and 50%. In this study, we compared two methods (traditional method and finite-resampling method) for estimating pharmacokinetic parameters. The traditional method (M1) directly parses the original data (that is, “single animal single-observation data” The finite resampling method (M2) is based on the observed time points of the original data permutations and combinations, it is expanded to a set of complete data containing the pharmacokinetic animal virtual data set composed of a new entity, the total 243 (C_3 ~ 1 × C_3 ~ 1 × C_3 ~ 1 × C_3 ~ 1 × C_3 ~ 1) Only virtual animals. In this study, a total of 100 repetitions were simulated. The simulation and parameter estimation were completed by NONMEM software. The results show that the CL and v estimated by the M2 method are closer to their corresponding simulation values, but slightly different in the combination of different IIV and residuals. Overall, the advantage of the M2 method decreases as the residual increases, and it also receives the influence of the II V size, and the M2 advantage decreases as the IIV increases. Similar to the M1 method, the M2 method does not have the ability to reduce the IIV of the parameters. Compared with the traditional method, the finite resampling method can provide more reliable results in estimating the pharmacokinetic parameters of “single animal individual - single observation type” data. Compared with the inter-individual variation, the estimation result mainly receives the influence of the residual size.